Auditor Tenure, Managerial Fraud, and Report Qualification: Theory and Evidence

Abstract

We consider the effect of auditor tenure on the level of managerial fraud and the extent of auditor qualification of reports. We consider two conflicting effects. As auditor tenure increases, the auditor's ability to detect fraud increases (the learning curve effect), which reduces the manager's fraud incentives. On the other hand, the auditor may become more sympathetic towards management (the loss of independence effect), which may increase fraud incentives. In order to analyse these issues, we develop an auditing game in which the manager makes an unobservable decision whether or not to commit fraud. The auditor then decides whether to perform a basic or extended audit. The level of audit affects the probability of fraud-detection. Following the outcome of the audit, the auditor then decides whether to issue a qualified or unqualified report. In equilibrium, the effect of auditor tenure on fraud and report qualification depends on whether the auditor's ability or the auditor's loss of independence dominates. Interestingly, longer auditor tenure may result in lower levels of report qualification due to two possible reasons; either an increase in auditor empathy (turning a blind eye to fraud), which suggests that auditor tenure may lead to more fraud, or an increase in the auditor's ability to detect fraud, which acts as a deterrence, leading to less fraud. Next, we consider evidence regarding the relationship between auditor tenure, managerial fraud and report qualification. First, we test our model using a database comprising a ten-year history of UK entities filing qualified financial statements in order to empirically test the relationship between length of auditor tenure and report qualification, and the impact of auditor change. We confine our analysis to simple qualification of financial statements. Our empirical results suggest that auditors with longer term relationships with their clients are less likely to qualify accounts in later years, and consider the impact of this on audit planning. Second, we conduct a survey of bankers, accountants and auditors in Pakistan regarding their perceptions of the relationship between auditor tenure and auditor independence. Finally, we develop and conduct a behavioural experiment regarding the effect of client-auditor relationships on report qualification. We conclude by considering the ethical dimensions of our model, and we provide policy implications in relation to the ongoing debate regarding mandatory turnover of auditors.